150 results about "Temporal consistency" patented technology

The invention discloses an intelligent report generation method and system based on multi-source heterogeneous data fusion, and particularly relates to the technical field of data fusion, and the method comprises the steps: constructing a fact anchor point containing a stable identifier in a fusion layer, and solidifying an aperture version signature and a time interval; packaging a screening condition, a sorting rule, an access plan, time window limitation and the like into a controlled context packet and freezing the controlled context packet into a unique fact source; in the range, generating sentences by using a limited template and writing sentence-level reference marks, so that each sentence can be traced back to a data slice and an evidence fragment; performing consistency verification according to a source, an aperture and time, performing conservative degradation on the attribution expression according to a consistency threshold value, and outputting a degradation strategy description; and finally, page two-way pointer and cross-sentence consistency recording and audit playback are realized through an evidence chain pointer table and a published version signature, and the problems of caliber drift, cross-window access and difficulty in evidence tracing are solved.

The invention relates to the technical field of mechanical arms, in particular to a man-machine interaction system of an intelligent mechanical arm with a body. The system decomposes a long time sequence task into sub-tasks through a large language model, and introduces a task acquisition module, a task queue management module, a task re-planning module and a control execution module, uses visual detection to identify gestures and environment events to trigger temporary tasks, and maintains interruptible marks and safety anchor points based on priorities and interruption risks. And during interruption, the mechanical arm with the timestamp and the environment state are collected, during recovery, the states are compared, local re-planning is carried out, transition sub-tasks are automatically generated for failure sub-tasks, and the mechanical arm is controlled to execute after constraint verification. According to the system, on the premise that the structure consistency, the space consistency and the time consistency are guaranteed, safe interruption and efficient recovery in a long-time-sequence task can be achieved, and the autonomy, the real-time performance and the operation safety of the mechanical arm in a man-machine cooperation scene are remarkably improved.

The invention discloses a cross-chain meteorological and hydrological data consistency verification method oriented to multi-block-chain collaboration, which effectively solves the problem of data alignment caused by inconsistent time standards in a multi-chain environment by collecting multi-source data and generating a data tuple with a geographic identifier and distributing a unified reference timestamp for the data in combination with a physical clock synchronization mechanism. And furthermore, the time deviation is extracted through a layered verification structure and inter-chain cross check, and the timestamp weight is dynamically adjusted, so that the time synchronization precision of the cross-chain data is improved. Finally, a standardized time sequence is generated through segmented aggregation of the data, cross-chain storage is completed, regular backtracking verification is performed in combination with a supervision chain, time consistency and content credibility of the meteorological and hydrological data in the cross-chain circulation process are guaranteed, and efficient collaboration and traceability between heterogeneous block chain systems are achieved.

The invention provides a multi-sensor fusion robot grabbing path planning method and device, and relates to the field of automation. The method comprises the following steps: constructing a first time correction relation based on observation parameters of a multi-source sensor; correcting the relation based on the first time; inputting the hardware timestamp synchronized to the master clock source into the extended state estimation model; iteratively updating the observation parameters of the multi-source sensor based on the extended state estimation model, and constructing a second time correction relation according to the updated observation parameters of the multi-source sensor; constructing a pose continuous time trajectory function corresponding to the target robot based on the second time correction relation; the continuous time trajectory function and the multi-source sensor observation parameters serve as input, and a grabbing path corresponding to the target robot is output through a time sequence consistency constrained joint optimization model; and controlling the target robot to execute grabbing operation based on the grabbing path. The problem that in the prior art, the precision of a robot grabbing path is low is solved.

The invention relates to the technical field of computer vision, in particular to a space-time consistency data generation method for visual target tracking. The method comprises the following steps: firstly, training a path generator on a target tracking training set, learning a motion law of a target in a time sequence by using optical flow estimation and conditional variation coding technologies, and generating a target motion track conforming to physical constraints; and then, based on the generated target trajectory, introducing a space-time consistency attention mechanism to guide a text-video generation model, and under the condition of keeping basic model parameter freezing, constraining the position, scale and continuity of a target in a generation frame through an attention network, thereby synthesizing a video frame sequence with real motion features. According to the method, target tracking video data with real motion characteristics and high time sequence consistency is generated, and the robustness of the model to complex motion, illumination change and shielding conditions can be improved in different scenes.

Systems and methods are disclosed related to a 3D grounded video foundation model. A video generation method and system provide 3D conditioning information to a video diffusion model to improve generated video quality (object and temporal consistency) that is grounded in three dimensions (3D). The video generation method and system also enable precise camera control, cinematic effects, and scene editing. Video output corresponding to a set of camera specifications is generated for a scene from input image(s) including one or more images of a static scene or a sequence of images (video) for a dynamic scene. The input image(s) are used to calculate a 3D cache representing the scene. The 3D cache is rendered according to the set of camera specifications to produce a frame sequence and a mask sequence that identifies missing pixels in each frame. The frame sequence is encoded and masked to generate the output video.

The invention discloses a method and a system for generating a long video with a controllable camera track based on a video diffusion model. Comprising a camera track and initial frame preparation stage, a point cloud construction and multi-view image generation stage, a scale factor alignment optimization stage, a camera motion prior injection and noise initialization stage, a diffusion inversion generation stage and a sliding window time consistency fusion stage. Three-dimensional camera track modeling, projection reconstruction and diffusion processes are combined, a video content generation process is explicitly guided to be aligned with a track path set by a user, and long video generation which is reasonable in structure, natural in vision and continuous in time is achieved.

The invention discloses a time-space consistency-based video super-resolution method and device for a complex operation scene, and a medium, belongs to the field of image and video processing, and aims at solving the problem that a stable, clear and continuous-structure video sequence is difficult to generate in an existing method, and the time-space consistency-based video super-resolution method for the complex operation scene based on video prior is provided. Comprising the following steps: utilizing submerged space modeling, mapping an input low-resolution video to a submerged space, and obtaining a corresponding submerged variable representation; initializing a random noise tensor with the same size as the latent variable expression as an initial noise state; in each step of sampling, dividing a noise state and latent variable representation into a plurality of blocks according to space and time dimensions; denoising is carried out on each tile block; the initial noise is converted into high-quality latent variable representation; and a high-resolution video is reconstructed. According to the method, the power inspection video with high resolution and high consistency can be generated, the generation stability is kept, and the video texture detail expression capability is improved.

The embodiment of the invention discloses a multi-process-based real-time multi-path camera synchronization method and system, according to the embodiment of the invention, frames with the same time are paired according to timestamps, real-time unified processing of multi-path incoming frames is realized through a shared memory circular buffer area, each camera acquires the frames in an independent process and attaches the timestamps, and the real-time synchronization of the multi-path cameras is realized. And interacting data through a shared memory circular queue, and dynamically matching frames according to a timestamp by a synchronization process. Compared with an existing scheme, data interaction real-time performance is improved through shared memory zero copy, synchronization precision is improved through timestamp matching and dynamic error control, expansibility is enhanced through a multi-process parallel framework, the precision and real-time performance of multi-camera data synchronization are greatly improved, and the method is suitable for large-scale popularization and application. The requirement of three-dimensional reconstruction for multi-view data time consistency is met through targeted optimization, high efficiency and accuracy of subsequent three-dimensional reconstruction and data processing are ensured, and synchronization precision and real-time performance are remarkably improved.

The invention relates to the technical field of trajectory control, in particular to a multi-modal motion trajectory real-time tracking method and system for machine control, which performs weight redistribution through channel noise variance and amplitude difference, performs weighted stacking and continuous updating in combination with Kalman filtering, reduces state estimation variance and channel bias transfer, and improves the tracking accuracy of a multi-modal motion trajectory. Time consistency and amplitude consistency of fusion track state data are improved, an overshoot peak value is inhibited, a deviation attenuation process is shortened, abnormal samples in a deviation sequence are rejected through an isolated forest, interference of outliers on parameter updating is reduced, and speed difference and acceleration difference are compared for a correction control sequence, so that the accuracy of correction control is improved. The boundary detection rate and the determinacy of the stage number are improved, the energy change rate is screened through the speed difference and the curvature value of the adjacent stages, the transfer relation is determined, the time sequence coherence and the geometric continuity of stage switching are kept, and the time sequence stability and the stage connection coherence of the control instruction are kept in the multi-degree-of-freedom space.

The invention relates to a coal mine flame pseudo fire interference suppression method based on time sequence consistency analysis, and belongs to the technical field of coal mine safety monitoring. According to the method, aiming at the problem of high false alarm rate of a flame detection system caused by false fire source interference such as electric welding sparks and mechanical friction in a mine environment, the flame identification accuracy is improved through time sequence consistency analysis. According to the technical scheme, the method comprises the following steps: acquiring an infrared spectrum flame saliency probability graph of continuous image frames in an underground coal mine; extracting a flame candidate region by adopting Otsu adaptive threshold segmentation in combination with morphological optimization and hole filling; and performing time sequence consistency analysis on the continuous frame candidate regions, and calculating the spatial overlapping degree, the morphological stability and the infrared thermal intensity change trend of adjacent frames. According to the method, the false fire false alarm rate is remarkably reduced, the robustness to complex backgrounds and dynamic illumination is enhanced, the method can be independently integrated to an existing infrared flame detection system or used in cooperation with a deep learning model, and the reliability of mine fire monitoring is improved.

The invention discloses an internal control risk dynamic identification method and system based on a multi-modal model, and relates to the technical field of risk dynamic identification methods. According to the method, through an unsupervised state transition analysis mode, the hidden abnormal jump paragraph in the behavior mode can be dynamically identified, manual rule setting is not needed, and the self-adaptive capability of risk identification is enhanced; when a time lag problem exists in a multi-modal log, through operation chain backtracking and a virtual calibration mechanism, the time consistency and causal integrity of a cross-modal behavior chain are ensured, and the misjudgment probability caused by record inconsistency is remarkably reduced; a state sequence is mapped into a process topological graph by adopting a sliding matching method, free fragments which cannot be matched structurally are effectively identified, and starting and ending points of risk behaviors can be accurately positioned.

The invention provides a clock synchronization method, a multi-axis drive control system and a computer storage medium, and the method comprises the steps: obtaining a clock synchronization signal which is sent by a master station timer and has a matched period at a slave station end, selecting any timestamp in the clock synchronization signal as a reference point, and carrying out the clock synchronization based on the reference point, determining the corresponding position of the reference point in the counting signal of the local timer, mapping the time information of the master station to a local time axis of the slave station, and calculating the phase difference between the corresponding position of the reference point in the counting signal of the local timer and the starting position of the next period of the counting signal of the local timer, and quantizing the real-time deviation of the master and slave clocks, and finally, dynamically adjusting and updating the period of a counting signal of a local timer to eliminate the phase difference, so that the slave station clock period is kept synchronous with the master station, and the time consistency of the action of each slave station in the multi-axis driving system is ensured.

The invention relates to the technical field of artificial intelligence, can be applied to business scenes of financial science and technology, medical health and the like, and discloses an animation sequence generation method, device, equipment and medium. Modeling space-time dependence in a dynamic graph attention module to generate a dynamic graph feature set, inputting the dynamic graph feature set into a time sequence feature enhancement module to obtain an enhanced feature set, performing up-sampling through an image decoding module to form a candidate animation sequence, generating a total loss item in combination with an adversarial loss item, a perception loss item and a time consistency loss item, and jointly updating all the modules to obtain the dynamic graph feature set. And outputting a target animation sequence after updating is completed. According to the method, through a dynamic graph attention mechanism and a time sequence enhancement mechanism, a cross-frame space relationship is adaptively modeled, long time sequence consistency is constrained, and end-to-end training is realized by combining multi-loss item joint optimization, so that distortion and jitter of a candidate animation sequence are effectively reduced, and stability is improved.

The invention discloses a video super-resolution reconstruction method and system based on artificial intelligence, and the method comprises the steps: firstly obtaining an original low-resolution video stream on a production line, carrying out the framing management of the video stream, arranging and storing, and guaranteeing the completeness of a data time sequence; the method comprises the following steps: preprocessing an original low-resolution video stream to obtain a clear and high-quality deblurred frame; performing feature extraction and spatial-temporal feature fusion according to the clear and high-quality deblurred frame, and performing reconstruction to obtain a high-resolution video frame; and according to the reconstructed high-resolution video frame, edge enhancement, time sequence consistency adjustment and color and brightness restoration are carried out to obtain a final super-resolution frame sequence. And finally, coding and synthesizing according to the final super-resolution frame sequence, synthesizing the single-frame sequence into a high-resolution video stream with a calibrated time sequence, and docking with a production line service system. According to the method, the quality and the time sequence integrity of original data are effectively improved, and the adaptive capacity of actual working conditions of a manufacturing site is remarkably enhanced.

This invention discloses a target trajectory association method, comprising: a preset area containing multiple detection devices, each detection device using polar coordinates as a coordinate system, reporting the first position information of the target, and each detection device acting as a pole, periodically reporting the first position information of the detected target; this target trajectory association method and system achieves temporal consistency of targets by periodically aligning two detection devices in the identification group, and unifies the state variables of targets reported by non-central devices to the coordinate system of the central device, achieving spatial consistency of targets. When the targets reported by the detection devices are consistent in both time and space, the similarity between targets reported by different detection devices in the same area is calculated to obtain the correlation between targets, thereby realizing the association of target trajectories.

The present invention relates to an event-driven Extract-Transform-Load processing system and corresponding method that enable real-time synchronization of data originating from multiple heterogeneous and asynchronously operating data sources. The invention implements a machine-implemented processing approach in which data events dynamically trigger transformation determination, synchronization execution, and multi-stage computational validation without reliance on predefined batch schedules. Transformation operations are adaptively selected based on contextual data characteristics, historical synchronization behavior, and dynamically adjustable precision thresholds, thereby ensuring accurate data alignment under evolving data structures and operating conditions. The system further performs sequential validation of structural conformity, temporal consistency, and cross-source relational integrity prior to permitting data persistence or downstream transmission.

A resource event sensing method and system based on computing network fusion relates to the technical field of computing network fusion resource sensing, and comprises the following steps: decomposing an application into a computing task and a connection task to generate application information; generating task information including a task ID, a task type, a resource type, a corresponding resource ID and resource context information for each task; constructing a unified computing network resource information topological graph based on the task information, taking computing resources corresponding to the computing tasks as nodes, and taking connection resources corresponding to the connection tasks as edges; a unified clock source is adopted to ensure the time consistency of sensing data, and the resource state is sensed; and carrying out aggregation and correlation analysis on the sensing events according to the resource dependency chain. Through the unified event model, the computing network topology and the resource association mechanism, the problems of sensing fragmentation, fault positioning difficulty and low resource scheduling efficiency caused by sensing separation of the computing power domain and the network domain are solved, a precise end-to-end resource view is provided for AI application, and the operation and maintenance efficiency of the AI cluster is remarkably improved.

An electronic control unit in which a plurality of workloads are identified from execution of a runtime application, where each identified workload is associated with a set of cache lines of a cache resource. A priority designation of the set of cache lines is determined based at least in part on a usage of the set of cache lines during an upcoming time interval. In examples, the usage is based at least in part on memory access operations that are expected to be performed by the plurality of workloads during the upcoming time interval.

This invention relates to the field of water resource optimization scheduling technology in water conservancy and hydropower engineering, specifically to a reservoir optimization scheduling method and related products based on intertemporal sensitivity sparse repair. The method involves constructing and solving implicit equations for power generation and water volume, establishing an evaluation system and calculating intertemporal sensitivity to obtain a set of key time periods. Then, sparse repair is implemented on the initial power generation plan within this set, and the repair operation is embedded in an iterative optimization framework. This invention avoids the use of empirical estimation or coarse trial calculations in traditional reservoir scheduling, enhances the water resource safety net of reservoirs, suppresses the oscillation phenomenon of traditional intelligent evolutionary algorithms, and maintains the scheduling intent and temporal consistency of the initial predetermined plan on the power grid side.

This invention discloses a laser-vision collaborative precision positioning method and system for column installation, comprising: acquiring the original laser point cloud sequence and monitoring images of the column installation base area; using a multi-dimensional confidence evaluation model to identify dust interference based on local density, reflection intensity, and temporal consistency to obtain a purified laser point cloud; identifying edge geometric features and coded identifiers in the monitoring images and performing sub-pixel level refinement; employing a multi-level pose alignment strategy, using the asymmetric constraints provided by the coded identifiers for initial pose estimation, and jointly optimizing the heterogeneous laser and vision data through an uncertainty weighted model to obtain the column installation pose; and driving a robotic arm to perform installation posture compensation based on pose deviation. This invention solves the positioning problem under dust interference and low-texture scenes, improves the stability and reliability of the positioning system, and realizes automated operation of column installation.

The invention provides an image-to-video representation migration method and device based on loop consistency and semantic constraint, and the method comprises the steps: carrying out the coding and mapping of video frames of an input video through an encoder and a mapping layer, constructing a loop consistency training target and the semantic separability constraint between videos, and forming an overall optimization target; the mapping layer is optimized through multi-round iteration; and obtaining the video representation of the target video through the encoder and the optimized mapping layer. According to the method, cross-frame correspondence is enhanced through circulation consistency, a position shortcut is broken, KL divergence regularization is introduced to keep separability, a set of lightweight projection and efficient migration overall scheme is designed, and under the condition that an encoder is pre-trained by freezing an image, the accuracy of projection is improved. And effective balance between the time sequence consistency in the videos and the semantic separability between the videos can be realized by a small amount of self-supervised training. Meanwhile, due to the fact that only the light projection layer is newly added, the method has the engineering advantages of being easy to implement, low in cost and wide in adaptation.

The invention discloses an automatic driving scene multi-view video generation method and system based on space-time consistency constraint. According to the method, in a diffusion model generation process, three attention mechanisms are integrated: based on cross-view-angle deformable cross attention of a camera pose, geometric consistency among different view angles is realized; on the basis of cross-frame cross attention of optical flow prior, motion coherence between adjacent frames is ensured; and on the basis of the global time sequence consistency constraint of cross attention, the reasonability of long sequence generation is ensured. According to the method, the problems of geometric distortion and time sequence breakage in multi-view video generation are effectively solved, high-quality dynamic scene videos with consistent time and space can be generated, and efficient data support is provided for automatic driving simulation.

This invention discloses a secure data transmission access system and method for intelligent medical IoT, relating to the field of data transmission control technology. It includes: S1, real-time acquisition of medical monitoring data and transmission information, and data preprocessing; S2, assessment of the temporal consistency of medical monitoring data sequences from each transmission channel, and data rearrangement and completion processing for transmission channels with sequence distortion; S3, analysis of the correlation of data change trends between different transmission channels, assessment of the time synchronization relationship between channels, and time shift alignment for channels with stable synchronization relationships; S4, construction of a patient physiological state representation, assessment of the coupling consistency between various physiological parameters, and determination of whether to allow subsequent patient state snapshot construction. This solves the problem of inaccurate physiological state representation caused by temporal disorder and cross-channel synchronization distortion during multi-source medical monitoring data transmission in an intelligent medical IoT environment.

The invention relates to the related technical field of deep learning, in particular to an efficient event stream spatio-temporal representation learning method based on progressive spatio-temporal token selection, and the method comprises the steps: analyzing the spatio-temporal consistency of event data through a spatial token purification module (STP), removing noise and a non-event region, evaluating the motion similarity between adjacent frames through a time sequence token compression module (TTC), and obtaining the spatial representation of the event stream. And a redundant motion mode is removed, so that the calculation overhead is remarkably reduced. The efficient event stream spatio-temporal representation learning method based on progressive spatio-temporal token selection is suitable for tasks such as action recognition and video anomaly detection of an event camera, experiments on HARDVS, DailyDVS-200 and SeACT data sets show that 28.8%-42.2% of calculation amount reduction can be achieved with slight precision loss, the reasoning speed is increased by 20.6%-41.3%, and the method has the advantages of being high in practicability, high in practicability and the like. The method has the advantages of light weight, plug and play and strong generalization.

This invention discloses a method for generating poses of articulated objects based on physical perception and graph diffusion, relating to the field of computer vision. The method first reconstructs the componentized geometric and physical properties of objects from images using a bi-branch neural implicit network and initializes component connectivity relationships. Subsequently, it refines the component relationship graph through kinematic fitting and temporal consistency checks, and uses a physically enhanced graph diffusion process to infer the prior pose distribution of components on an SE(3) manifold. Finally, using this prior and reconstructed information as conditions, a conditional diffusion model is constructed on the SE(3) manifold, and backsampling is performed through a physically guided two-step backsampling framework to generate a diverse and physically plausible set of pose assumptions for articulated objects. This invention achieves efficient generation of diverse and highly physically plausible poses of articulated objects by tightly coupling physical laws with data-driven generation.

The application discloses a remote control and data interaction method and system of a skid loader based on edge computing, and relates to the technical field of intelligent control of engineering machinery, and comprises the following steps: in view of the data interaction of multiple edge nodes in the remote control process of the skid loader, the sampling period and the instruction sending period of each edge node are sorted respectively, a unified time reference is established, and a time sequence is formed. Through the establishment of the unified time reference and the continuous time sequence, the time consistency between the multiple edge nodes is realized, the instruction flow fault is avoided, and the smooth transition of actions is ensured; meanwhile, through the construction of a cache misplacement comparison table and the re-planning of the data refresh sequence, the cache update is synchronized, the instruction transmission is continuous, and therefore the response continuity and the control stability of the remote control of the skid loader are improved.

The invention provides a deep counterfeit face detection method fusing time sequence consistency and space consistency, and belongs to the field of counterfeit detection, and the method comprises the steps: obtaining a target video; for each video frame containing the face in the target video, executing a space-time fusion feature extraction operation to obtain a space-time fusion feature corresponding to the video frame; determining whether the target video is a forged video based on the space-time fusion feature corresponding to each video frame; wherein the space-time fusion feature extraction operation comprises the steps of determining a space consistency feature corresponding to the video frame based on the video frame, and determining a time sequence consistency feature corresponding to the video frame based on the video frame and a previous video frame of the video frame; and performing space-time fusion by taking the space consistency feature corresponding to the video frame as a query vector and taking the time consistency feature corresponding to the video frame as a key vector and a value vector to obtain a space-time fusion feature corresponding to the video frame. According to the invention, the accuracy and reliability of deep fake face detection can be improved.

This invention discloses a temporally enhanced knowledge graph-driven dialogue agent and its implementation method, belonging to the fields of artificial intelligence and natural language processing technology. The dialogue agent includes a dialogue interface, natural language understanding, a temporal knowledge graph, a temporal reasoning engine, dynamic update and failure management, and a natural language generation module. It includes designing the basic representation unit of the knowledge graph as a temporal quintuple; designing a temporal reasoning engine and a dynamic update and failure management method; designing a method for matching, verifying, and performing multi-hop reasoning of temporal information in natural language problems with temporal facts in the knowledge graph; achieving accurate parsing of temporal constraints in dialogue, real-time adaptation of dynamic knowledge, rigorous temporal logic reasoning, and continuous maintenance of contextual temporal consistency; and outputting a temporally accurate, knowledge-reliable, and logically coherent intelligent response, capable of meeting the needs of complex scenarios such as vertical industry decision support and high-reliability intelligent interaction.

The application relates to the technical field of power distribution systems, in particular to a power distribution system intelligent collaborative decision-making method, which comprises the following steps: S1, collecting the sampling frequency, uploading delay and data integrity identification of each sub-region sensing node, analyzing the consistency of operation data structure, identifying time alignment defects or logical coupling faults, and generating a heterogeneous drift feature map; S2, combining historical data, topological mechanisms and boundary rules, reconstructing control input, and generating a data consistency control input set; S3, based on the input set, scheduling targets and system constraints, constructing and distributing a collaborative strategy, and realizing intelligent collaborative regulation and control. According to the application, the heterogeneous drift feature map is constructed, the spatio-temporal consistency operation data is reconstructed, and the collaborative strategy set with adaptive execution capability is generated, so that the accurate identification, effective repair and intelligent collaborative regulation and control of the power distribution system under the condition of inconsistent data are realized.